search

geneontology / go-shapes

Schema for Gene Ontology Causal Activity Models defined using RDF Shapes
2 stars 0 forks source link

build the validation test cases #8

Open goodb opened 5 years ago

goodb commented 5 years ago

We need to assemble a collection of go-cam models (as ttl files) that correctly validate and another collection that fails (for predictable reasons) so that we can validate the validation code.

Anyone that has been working on developing the specification document would be a good candidate for helping to build these exemplars. Ping @lpalbou

pgaudet commented 5 years ago

Sure !

goodb commented 5 years ago

@pgaudet if you are leading the charge on this. Could you go ahead and put your examples in the /test_ttl/go_cams/should_fail and /test_ttl/go_cams/should_pass folders ? (or reorganize as you see fit).

I put some examples together to get the ball rolling. https://github.com/geneontology/GO_Shapes/commit/5a3d973ea84104fe477a278692135608b075ac77

Would be awesome to see such files explicitly used and referenced from the specification document to illustrate intended semantics. I have so far been making them and storing them here in this repo but, if they became more important as documentation we could put them in the main noctua model collection such that they could be linked to and browsed easily.

vanaukenk commented 5 years ago

@goodb

Here are some links to models that would be good to test:

http://noctua.geneontology.org/editor/graph/gomodel:5b528b1100000186 (this is still development, but should fail our current specs)

http://noctua.geneontology.org/editor/graph/gomodel:568b0f9600000284 (this one is production and should pass our current specs)

Do you need me to supply the ttl files or can they be retrieved from the gomodel id?

goodb commented 5 years ago

thanks @vanaukenk . Could you let me know why the first should fail?

No problem getting the ttl (the 'export OWL' option in Noctua will do that via URLs like: http://noctua.geneontology.org/download/gomodel:5b528b1100000186/owl )

vanaukenk commented 5 years ago

@goodb

The first model should fail because it has MFs that are 'part of' other MFs.

pgaudet commented 5 years ago

Hi @goodb

test cases are being assembled here: https://docs.google.com/spreadsheets/d/1ksUFbU1OBa02U1Nc-2A0hFzGfE3MYIMHXmUReWUYTJs/edit#gid=0

Let me know if this works for you.

Thanks, Pascale

goodb commented 5 years ago

This looks good to me, but passing the torch for the moment to @kltm @balhoff @cmungall and hopefully also @lpalbou .. Note there is the possibility of losing these models if they are newly created on dev. For things we want stable, my practice would be to put them on master with the appropriate tag - e.g. development , or maybe we should add a new tag 'demo'.